Self-report diary: a method to measure use of office lighting.
Electricity constitutes half of the energy use in nonresidential buildings in Sweden, and lighting accounts for the largest percentage, about 20 percent [The Swedish Energy Agency, 2007]. Obviously, electricity use for lighting is affected by the length of time the lighting is switched on [Dubois and Blomsterberg, 2011]. The most common strategies for reducing energy use at work places are structural and operational changes such as installing energy efficient lighting [Starik and Marcus, 2000]. However, employees' energy use behaviors have become an area of increasing interest [Scherbaum and others, 2008]. Office occupants ' behavior has a significant effect on energy wasted during nonwork hours. Therefore, behavioral change has potential to save energy, which could be higher than that of technical solutions [Masoso and Grobler, 2010]. Behavioral change can be encouraged by, for example, an awareness campaign to remind the employees about dimming and turning the light off [Galasiu and others, 2007]. Changes in employees' energy use behaviors are possible to decrease organizational energy use can be taken away [Siero and others, 1996] and are needed to support structural or operational changes [Scherbaum and others, 2008]. Further, a better understanding of the relationship between behaviors and psychological factors in relation to lighting will facilitate the selection of lighting choices in work places [Veitch and Gifford, 1996].
A wide range of methods have previously been used to study the effects of office occupant behavior on lighting energy use; for example, interviews [Escuyer and Fontoynont, 2001], observations using time-lapse photography [Hunt, 1979], observation by humans [Boyce, 1980; Maniccia and others, 1999], video surveillance [Reinhart and Voss, 2003], and use of data loggers with sensors capturing motion, illuminance, energy use and other related activities [Love, 1998; Reinhart and Voss, 2003; Galasiu and others, 2007; Masoso and Grobler, 2010]. Although these methods are widely used, Boyce  points to office occupants' reluctance to accept the use of electronic recording and time-lapse photography and to participate in interviews. In addition, methods such as data logging or video surveillance may be too expensive when a large sample size is required. Moreover, using a video camera to record behavior may affect the occupants' attention to the behavioral aspects of the study [Reinhart and Voss, 2003].
Self-reports are commonly used to measure human behavior since they are inexpensive, easy to administer and do not impinge on the privacy of the participant [Bell and others, 2001]. The self-report diary has been applied to capture energy use patterns in households and has been proposed as a cost effective approach to measure energy use [Tso and Yau, 2003]. Palm and Ellegard , for example, used a time-geographical diary and self-reported data to identify and visualize households' use of electrical appliances as well as their need for lighting and heating. However, in offices, there has been little effort to identify the basic behavioral patterns of occupants in relation to the use of electric lighting [Reinhart and Voss, 2003]. If the self-report diary can capture as well as monitor the long-term behavior of office occupants, this straightforward and inexpensive method could contribute to a better understanding of lighting and energy use in offices. Moreover, field study data obtained by the diary could help to develop and use simulation techniques to refine the occupants' behavioral patterns.
To the best of the authors' knowledge, self-report diaries have not been used previously to measure occupants' use of office lighting. However, before the self-report diary can be used to investigate such behavior, certain questions about the viability of the method need to be studied. Self-report diaries in households have been shown to be highly dependent upon the extent of the participants' involvement in the study, and according to Horelli [2002, p 616], people are sometimes not motivated to participate. In addition, the validity of self-reports has been questioned [Gatersleben and others, 2002; Chao and Lam, 2011]. Self-reported data regarding occupant use of manual lighting controls in private offices revealed that people do not always report their own actions accurately, which leads to the need to use supplementary objective measurement [Maniccia and others, 1999]. Chao and Lam  found that although self-reports related to actual behavior, people may overestimate the frequency of their pro-environmental behavior such as turning off table lamps, shutting down computers, and so on. Consequently, researchers should be careful when interpreting results based only on self-reports, and should strive to use multiple modes of measures. A comparison of diaries with meter reading or electronic record would provide an opportunity to examine the self-report against other types of measuring techniques which requires more research resources.
2 AIM AND OBJECTIVES
The aim of this paper is to examine the suitability and reliability of diaries based on self-reported behavior to measure occupants' use of office lighting and energy as compared to data from electronic measurement. The specific objectives were:
(i) To examine data obtained by a self-report diary to study office occupants' behavior in relation to lighting use.
(ii) To examine relationships between occupant-behaviors, occupancy and light-on time by comparing data obtained by the diary and those obtained by electronic measurement.
(iii) To examine the use of diaries for short- and long-term studies.
(iv) To compare data on energy use obtained by the diary with those measured electronically regarding the length of light-on time and occupancy.
This paper is part of a larger research project studying occupants' experience and use of different lighting controls carried out at Lund University, Department of Architecture and Built Environment.
3 METHOD AND MATERIALS
The study was conducted in 18 single-occupant offices where the occupants had individual control of their office-spaces. The offices had an approximate area of from 12.00 [m.sup.2] to 20.00 [m.sup.2] with 2.40 m of floor-to-ceiling height and were all located in an office building in Stockholm, Sweden. All offices had access to natural light through two-or four-glazed windows equipped with window blinds and a layer of fabric curtains that could be operated manually. Electric lighting in each office was provided by a manually or automatically controlled ceiling luminaire and by a desk lamp that could only be controlled manually. The luminaire was suspended over the center of the desk at about 0.5 m below the ceiling. The different lighting controls for the luminaires are shown in Table 1.
TABLE 1. Lighting controls of ceiling luminaires in the offices Office Lighting Control Manual Control Device Office 1 Manual on/off A wall-mounted button switch located close to the door Office 2 Office 3 Office 4 Manual dimming and on/off A pull cord switch located under the luminaire Office 5 Office 6 Office 7 Manual on/off, and automatic A pull cord off after 15 min vacancy switch located under the luminaire Office 8 Office 9 Office 10 Manual on/off, and automatic A pull cord lowering of lighting level to switch located 90% of lighting setting of 500 under the lx and off after 15 min luminaire vacancy Office 11 Office 12 Office 13 Automatic on, lowering of None lighting level to 80% of lighting setting of 500 lx after 15 min vacancy and off after 2 h vacancy Office 14 Office 15 Office 16 Manual on/off with automatic A wall-mounted on/off option, automatic off push button after 30 min vacancy, and switch located algorithmic control increasing close to the lighting level at 8.00-10.00 door and 13.00-14.00 Office 17 Office 18
Eighteen office occupants were approached at a meeting and were informed of the study's background, aims and procedures and that participation in the self-reported part was voluntary. They were also informed about electronic measurement of their occupancy and use of ceiling luminaires by data logger throughout the entire study period and that both self-reported- and logged data were treated as confidential. However, they were not officially informed about the comparisons between self-reported- and logged data. All occupants were employed by a property development and management company. They worked in their respective offices during the study period from June 2009 to the end of May 2010.
The occupants signed a written consent to participate. Four of the original occupants dropped out of the study and one was replaced in one office by a new occupant who participated from February 2010. In total, 15 office occupants (six males and nine females, mean age[euro]47, age range: 29-62 years) participated in the self-reported part. They were familiar with the lighting controls in their offices before the study was conducted.
3.3.1 DIARY FORM
The diary form in Fig. 1 was used to capture the occupants' lighting-use-related activities in each office during a workday. On the form's front page, the occupants were asked to specify the date on which they completed the diary form.
HOW I USE LIGHTING IN MY OFFICE WHEN I WORK Fill in all lighting-use-related activities Activities Movement Time Ceiling Desk Sun Lonation lamp lamp blocking * Switch * Increase * Switch * 100% * Sitting in my on lighting on office level * Switch * Switch * 50% * Leaving my office off off but staying in my department * Do * Lower * Do * 25% * Leaving my office nothing lighting nothing for another level department/floor * 0% * Going outside the office building * Switch * Increase * Switch * 100% * Sitting in my on lighting on office level * Switch * Switch * 50% * Leaving my office off off but staying in my department * Do * Lower * Do * 25% * Leaving my office nothing lighting nothing for another level department/floor * 0% * Going outside the office building * Switch * Increase * Switch * 100% * Sitting in my on lighting on office level * Switch * Switch * 50% * Leaving my office off off but staying in my department * Do * Lower * Do * 25% * Leaving my office nothing lighting nothing for another level department/floor * 0% * Going outside the office building * Switch * Increase * Switch * 100% * Sitting in my on lighting on office level * Switch * Switch * 50% * Leaving my office off off but staying in my department * Do * Lower * Do * 25% * Leaving my office nothing lighting nothing for another level department/floor * 0% * Going outside the office building * Switch * Increase * Switch * 100% * Sitting in my on lighting on office level * Switch * Switch * 50% * Leaving my office off off but staying in my department * Do * Lower * Do * 25% * Leaving my office nothing lighting nothing for another level department/floor * 0% * Going outside the office building * Switch * Increase * Switch * 100% * Sitting in my on lighting on office level * Switch * Switch * 50% * Leaving my office off off but staying in my department * Do * Lower * Do * 25% * Leaving my office- nothing lighting nothing for another level department/floor * 0% * Going outside the office building Fig. 1. The diary from.
Each occupant was asked to report on the following set of activities: regulation of a ceiling luminaire: (i) switch on, (ii) switch off, and (iii) do nothing (addition-ally, if relevant: (iv) increase and (v) lowering of lighting level with manual dimmers); regulation of a desk lamp: (i) switch on, (ii) switch off, and (iii) do nothing; adjustment of window blinds together with curtains to block daylight from the office's windows: 100 percent blocked, 75 percent blocked, 50 percent blocked and 0 percent blocked. The occupant was also asked to report activities together with different movements: (i) sitting in the office, (ii) leaving the office but staying in the department, (iii) leaving the office for another department/floor and (iv) going outside the office building. She or he was also asked to indicate at what time of day each reported activity with movement was performed. By documenting the times when the activities and movements were reported, the length of light-on time and occupancy time could be calculated.
3.3.2 DATA LOGGER
A data logger measured the occupants' use of ceiling luminaires and presence in their offices (occupancy). PIR sensors installed in the offices were connected to an AHLBORN's ALMEMO[R] measuring instrument which was connected to a software program AMR WinControl V6 in a laptop computer. This data logger detected and recorded the electric current produced by the ceiling luminaires (represented in amperes, [greater than or equal to] 0.1 if the luminaire was on) and occupancy (represented as 100 percent for occupancy and 0 percent for vacancy) in each office. The logger was set to detect and record the data every two minutes. The data from the electric current were used to calculate the length of light-on time; and the data for occupancy-vacancy were used to calculate occupancy time.
3.4.1 DATA COLLECTION AND RESPONSE RATE
During the study period, the occupant of each office was asked once every two months to fill in the diary form during a workday. Originally, we expected that six diary forms from six workdays would be submitted by each occupant. However, for offices 16, 17 and 18, the three occupants were only able to submit the forms for five workdays due to a technical problem with lighting controls in their offices. The data from the data logger were collected throughout the entire study period from June 2009 to the end of May 2010.
A total of 105 diary forms were distributed and 59 diary forms were returned, with four extra returned by one occupant. One diary form was returned without data and four filled-in diary forms by one occupant were eliminated due to errors in the electronic measurement of occupancy in his office. In total, 54 diary forms reported by 14 occupants were used for the data analysis. Of the 54 diary forms, 31 forms by six occupants were filled in for every occasion on which they were asked to report, whereas 23 forms by eight occupants were filled in occasionally, that is, on from one to four occasions.
3.4.2 DATA ANALYSIS
To examine the suitability of the diary form for measuring the occupants' use of office lighting, we first focused on data for lighting-use-related activities that were expected but were missing from the returned diary forms. Then, the self-reported activities were examined for self-reported activities regulating electric lighting and the diary forms were categorized according to the types of lighting controls such as fully manual (manual on/off and manual dimming), manual on/auto off, and manual control with auto on/off option.
To examine the suitability and reliability of the diary form, the self-reported data was compared against the electronic measurements. We focused on comparing differences between self-reported- and logged data for four variables: (i) activities regulating ceiling luminaires, (ii) movements, (iii) light-on time and (iv) occupancy time. We categorized the occupants into two groups in accordance with the extent to which they participated in the self-reported part: full participation occupants were the six who completed 31 diary forms, and occasional participation occupants were the eight who completed 23 diary forms.
Data for the two variables reported in diary form (i) activities regulating ceiling luminaires, and (ii) movements were compared against the data recorded by the logger on the same dates. The self-reported activities were compared to the logged data of electric current. The self-reported movements were categorized either as occupancy or vacancy and then compared to the logged data for occupancy-vacancy. We examined the data from the diary forms and classified them as either correct, incorrect or missing data entries. We considered data to be missing when an activity or movement was not reported in the diary form within about 10 minutes after they were recorded in the logger. Data were considered correct when the activity or movement was reported correctly at the same time or within 10 minutes of being logged. This was done to allow for any time differences between individuals' watches and the logger. Relationships between the self-reported- and logged data were analyzed using nonparametric correlation analyses, Spearman's [rho] (Pallant 2007).
Data for (iii) light-on time and (iv) occupancy time were calculated in hours using the self-reported- and logged data. The relationships of these two variables were analyzed using parametric correlation analyses, Pearson's r. To further examine the reliability of the diary form, we analyzed the relationships between light-on time and occupancy time in the self-reported- and logged data, and then comparing the data from these two sources. To examine the suitability and reliability of the diary form for measuring lighting energy use, we tested by paired-samples t-test with a two-tailed test of significance whether there was any difference between the mean values of the self-reported- and logged data for these two variables.
Missing data in all correlation analyses were treated with the Exclude cases pairwise option. All analyses were performed by IBM SPSS Statistics 19.
Of the 54 diary forms, 41 forms (76 percent) reported all of the lighting-userelated activities. Self-reported dimming was reported very few times in the six diary forms by one occupant. As shown in Table 2, data for movements were reported in all diary forms, whereas data for other activities were not reported in all the forms.
TABLE 2. Number of the diary forms obtaining self-reported data for lighting-use-related activities Lighting-Use-Related No. of the Diary Forms Activities with Self-Reported Data Movements 54 (all) Activities regulating a 53 ceiling luminaire Activities regulating a 51 desk lamp Activities adjusting 41 window blinds and curtains
4.1 SELF-REPORTED BEHAVIORS CAPTURED BY THE DIARY FORM
4.1.1 THE REGULATING OF ELECTRIC LIGHTING
The regulating of ceiling luminaires and/or desk lamps was reported in 51 diary forms. In 14 forms of five occupants, desk lamps were reported to have been used together with ceiling luminaires. In five forms of two occupants, it was reported that they only used desk lamps, whereas in the remaining 32 forms, occupants reported that only ceiling luminaires were used. Self-reported activities regarding regulating ceiling luminaires and desk lamps together with movements are presented for all diary forms in Table 3.
TABLE 3. Self-reported activities regulating ceiling luminaires and desk lamps together with movements can be taken away Number of Times the Activity Reported in the Diary Forms Ceiling Luminaires Activity Fully Manual on/ Manual Desk together with Manual Auto Off Control Lamps (19 movement in the Controls Controls (13 with Auto Diary office (23 Diary Diary Forms) On/Off Forms) Forms) Option (8 Diary Forms) Manually 21 12 8 19 switched on when arriving in the morning Left the light 64 32 41 41 on when leaving during the day Manually 18 1 0 3 switched off when leaving during the day Manually 19 19 5 5 switched on when sitting during the day Did nothing when 6 13 - 0 sitting during the day (the light was off) Did nothing (the 64 16 34 27 light was already on) when sitting during the day Manually 22 10 6 15 switched off when leaving at the end of the day Let the light be - - 6 - on automatically when sitting during the day
In most of the diary forms, occupants reported that they switched on ceiling luminaires manually when they arrived in their offices in the morning and that they switched off the luminaires manually at least once, that is, at the end of the day. In 23 diary forms, occupants with the fully manual controls reported that they usually left on the ceiling luminaires when they left their offices (64 times), and when they returned (64 times). There were 18 times that the occupants switched off the luminaires when they left and 19 times that they switched them back on when they returned to their offices.
In 21 diary forms, occupants with the manual on/auto off controls and those with manual control with auto on/off option reported that they usually left the manual controlled luminaires on when they left their offices during the day. Then, the automatic controlled luminaires were switched off after vacancy but there was one time that the automatic controlled luminaire was switched off manually. The manual controlled luminaires were switched on when the occupants returned to their offices (24 times in total). In eight out of the 21 forms, there were six times reported by occupants with manual control with auto on/off option that the luminaires had been left on automatically when the occupants returned to their offices.
Generally, the occupants switched on manual controlled ceiling luminaires once when arriving in the morning and switched them off once when leaving at the end of the day. Some occupants switched off the luminaires when they left for a long period during the day, mostly for lunch time. There was no difference in self-reported switching behavior between the two types of control devices: wall mounted switch and pull-cord.
Similar to the use of ceiling luminaires, desk lamps were usually switched on upon arrival in the morning and switched off upon departure at the end of the day. In 19 diary forms, occupants reported that desk lamps were mostly left on when they left during the day (41 times) and returned to sit in their respective offices (27 times). There were a few times that the lamps were switched off when the occupants left (three times) and switched on when they returned to the offices (five times).
4.1.2 THE USE OF DAYLIGHT
The use of daylight by adjustment of window blinds and curtains was reported in 41 diary forms. Ten occupants reported in 37 diary forms that they did not adjust window blinds and curtains during the day. Regarding this, 23 diary forms (62 percent) reported that full daylight was allowed into the offices, 12 diary forms (32 percent) reported that 50 percent daylight was allowed. Two diary forms reported that 75 percent daylight and no daylight were allowed, respectively. Two occupants reported in another four diary forms that they adjusted window blinds and curtains when they sat in their offices.
4.2 RELATIONSHIP BETWEEN SELF-REPORTED- AND LOGGED DATA
4.2.1 ACTIVITIES REGULATING CEILING LUMINAIRES
The occupants' self-reported data were missing for 38 percent of the responses on regulating of ceiling luminaires (N = 466) compared to the logged data for the same activity (N = 752). The remaining 62 percent of self-reported data were correct for 53 percent and incorrect for 9 percent.
Table 4 shows that the self-reported- and logged data correlated strongly with each other. For the two groups of occupants classified according to extent of participation, the correlations were also strong. Considering the group with full participation, the correlations for the first three occasions were strong, whereas the correlations for the last three occasions were between strong and moderate. All the correlations were significant at the 0.01 or 0.05 level.
Self-reported data were missing for 36 percent of the responses regarding occupancy-vacancy (N = 468) compared to the logged data for the same activity (N = 735). The remaining 64 percent of self-reported data were correct for about 50 percent and incorrect for 14 percent.
Table 4 shows that the self-reported- and logged data correlated strongly with each other. The correlation for the group with full participation was strong, whereas the correlation was moderate for the other group. Considering the group with full participation, the correlations for the first three occasions were considerably stronger than the correlations for the last three occasions. All the correlations were significant at the 0.01 or 0.05 level.
TABLE 4. Spearman's p between self-reported- and logged data for activities regulating ceiling luminaires and for occupancy-vacancy Activities Occupancy-Vacancy Regulating Ceiling Luminaires Group of N Correlation N Occupants coefficient (p) Occupants with 280 0.65 ** 267 full participation (31 diary forms) 1st occasion (6 65 0.66 ** 51 diary forms) 2nd occasion (6 69 0.76 ** 71 diary forms) 3rd occasion (5 51 0.97 ** 51 diary forms) 4th occasion (5 34 0.45 ** 33 diary forms) 5th occasion (5 32 0.66 ** 32 diary forms) 6th occasion (4 29 0.41 * 29 diary forms) Occupants with 186 0.60 ** 201 occasional participation (23 diary forms) All occupants 466 0.62 ** 468 (54 diary forms) Group of Correlation Occupants coefficient (p) Occupants with 0.66 ** full participation (31 diary forms) 1st occasion (6 0.71 ** diary forms) 2nd occasion (6 0.82 ** diary forms) 3rd occasion (5 0.75 ** diary forms) 4th occasion (5 0.39 * diary forms) 5th occasion (5 0.50 ** diary forms) 6th occasion (4 0.54 ** diary forms) Occupants with 0.46 ** occasional participation (23 diary forms) All occupants 0.57 ** (54 diary forms) ** p < 0.01, * p < 0.05.
4.2.3 LIGHT-ON TIME
Table 5 shows a strong, significant correlation between overall light-on time calculated using the self-reported data and the time calculated using the logged data. For the two groups of occupants, the self-reported- and logged data also correlated strongly and significantly with each other. For the group with full participation, the correlations were very strong and significant for the first two occasions, whereas the correlations for other occasions were nonsignificant. All the significant correlations were at the 0.01 level.
TABLE 5. Pearson's r between self-reported-and logged data for light-on time and for occupancy time Group of Light-On Occupancy Occupants Time Time During During a a Workday Workday N Correlation N Correlation Coefficient Coefficient (r) (r) Occupants with 30 0.61 ** 31 0.26 full participation (31 diary forms) 1st occasion (6 6 0.95 ** 6 0.70 diary forms) 2nd occasion (6 6 0.92 ** 6 -0.28 diary forms) 3rd occasion (5 4 0.81 5 0.56 diary forms) 4th occasion (5 5 0.41 5 0.15 diary forms) 5th occasion (5 5 0.69 5 0.46 diary forms) 6th occasion (4 4 0.66 4 0.21 diary forms) Occupants with 22 0.76 ** 22 0.48 * occasional participation (23 diary forms) All occupants 52 0.71 ** 53 0.33 * (54 diary forms) ** p < 0.01, * p < 0.05.
4.2.4 OCCUPANCY TIME
As shown in Table 5, there was a moderate, significant correlation between overall occupancy time calculated using the self-reported data and that using the logged data. The correlation for the group with full participation was nonsignificant, for every occasion. The correlation for the occasional participation group was moderate and significant. All the significant correlations were at the 0.05 level.
4.3 RELATIONSHIP BETWEEN LIGHT-ON TIME AND OCCUPANCY TIME
Table 6 shows the correlations between light-on time and occupancy time calculated using the self-reported- and logged data, respectively. Overall, light-on time and occupancy time calculated using the self-reported data correlated moderately and significantly with each other. The correlation was weak and nonsignificant for the group with full participation, whereas there was a strong, significant correlation for the other group. In contrast, all correlations between light-on time and occupancy time calculated using the logged data were weak and nonsignificant.
TABLE 6. Pearson's r between light-on time and occupancy time Source of Data for Group of N Correlation Light-On Time and Occupants Coefficient Occupancy Time (r) Self-reported Data Occupants with 30 0.21 full participation (31 diary forms) Occupants with 21 0.58 ** occasional participation (23 diary forms) All occupants 51 0.42 ** (54 diary forms) Logged data Occupants with 31 0.29 full participation (31 dates) Occupants with 22 -0.12 occasional participation (23 dates) All occupants 53 0.08 (54 dates) ** p < 0.01.
4.4 MEAN VALUES FOR LIGHT-ON TIME AND OCCUPANCY TIME BASED ON SELF-REPORTED- AND LOGGED DATA
4.4.1 MEAN VALUES FOR LIGHT-ON TIME
There was no significant difference in mean values for light-on time (hours) calculated using all the self-reported- (M=7.00, SD=3.25) and logged data (M=6.68, SD=2.94), t (51)=-0.99, p=0.33. This result was true both for occupants with full participation (self-reported- (M=7.00, SD=2.48) and logged data (M=6.81, SD=2.02), t (29)=-0.64, p = 0.53), and for occupants with occasional participation (self-reported- (M[euro]7.00, SD=4.13) and logged data (M=6.51, SD=3.92), t (21)=-0.75, p=0.46).
4.4.2 MEAN VALUES FOR OCCUPANCY TIME
There was no significant difference in mean values of hours calculated using all the self-reported- (M=5.29, SD=2.20) and logged data (M=4.77, SD=1.45), t (52)=-1.70, p=0.09. For the group with full participation, there was also no significant difference in mean values from the self-reported- (M=4.69, SD=1.86) and logged data (M=4.68, SD=1.66), t (30)=-0.02, p=0.99. In contrast, there was a significant difference in mean values from the self-reported- (M=6.12, SD=2.41) and logged data (M=4.89, SD=1.13), t (21)=-2.73, p=0.01 for the group with occasional participation.
5.1 SUITABILITY OF THE DIARY FORM
The diary forms were successful in obtaining useful data regarding the occupants' use of ceiling luminaires, desk lamps, and movements. The data were missing from only a few of the returned forms. On the other hand, the suitability of the diary form for obtaining data regarding the use of window blinds and curtains can be questioned, since the data were missing from about one-fourth of the returned forms.
In the present study, self-reported data was able to reveal the occupants' use of electric lighting together with movements in and out of their offices. Most occupants usually switched on the lights manually upon arrival in the morning. Rarely, did they switch the lights off manually upon departure during the day, especially when they left the offices for short periods. However, they switched the lights off manually at least once, that is, at the end of the day. These switching patterns are similar to those found in previous studies using different techniques to observe lighting use behavior in offices [Jennings and others, 2000; Reinhart and Voss, 2003; Moore and others, 2003; Boyce and others, 2006]. Consequently, the diary form seems suitable in at least two respects (i) to assess the use of electric lighting during the day and, (ii) to determine behavioral patterns regarding the use of lighting.
The percentage of incorrect data entry for self-reported occupancy-vacancy (14 percent) was 5 percent higher than for self-reported data for activities regulating the luminaires, whereas the percentages of missing data entry were similar (36-38 percent) for both occupancy-vacancy and for activities regulating luminaires. Comparison with the logged data appeared to show that some occupants reported earlier arrival times and later departure times for their workdays. The logged data showed that missing data entries in the diary forms were underreported for activities of short duration. It is likely that the occupants did not report activities when they left for the toilet or printer. Also, the data logger may have detected other employees entering the offices to deliver documents or clean.
Most correlations were significant at the 0.01 or 0.05 level for the self-reported- and logged data concerning the variables (i) activities regulating ceiling luminaires and (ii) light-on time. These correlations indicate that the diary form can be used to measure office occupants' use of general lighting (ceiling luminaires). As for its use for long-term studies, the correlations of light-on time for the occupants with full participation were very strong and significant only for the first two occasions. This finding may indicate that the more occasions of self-reporting, the less accurate the data for the activities and the time reported by the occupants. Therefore, this result suggests that the reliability of the diary form for measuring light-on time decreases with the number of times an occupant is asked to complete the form.
Considering the occupants' time of occupancy of the offices, all correlations between self-reported- and logged data of occupancy-vacancy were significant. However, the correlations for occupancy time were rather weak, particularly for the occupants with full participation. When the relationship between light-on time and occupancy time was examined, no significant correlation with the logged data was found for these two variables. In contrast, a significant correlation at the 0.01 level was found for the self-reported data. These results could possibly have been affected by the number of incorrect or missing data entries for movements; and therefore suggest that the diary form could be unreliable for measuring occupancy time.
Olsen  suggests that self-reported behavior is rather a reflection of individuals' perception and beliefs about their own behavior than of actual behavior. Some occupants may report their perception and beliefs about the use of ceiling luminaires upon occupancy-vacancy rather than actual behaviors. Moreover, the occupants may rather report their habits. This response is due to the habitual nature of switching on lighting at the work place and, possibly, to that it is a sign of arrival at work [Moore and others, 2003]. In conclusion, the occupants more accurately reported their use of ceiling luminaires than their presence in their respective offices. Taken together, the results suggest that the suitability of the diary form for measuring office occupants' presence could still be questioned.
5.2 ENERGY USE
Overall, there was no significant difference in mean values for light-on time between the self-reported- and logged data. However, the results show that the mean values obtained from the self-reported data was slightly higher than for the ones from the logged data. These results thus suggest that people may tend to overestimate their light-on time when using the diary form.
Mean values of occupancy time, in comparison with mean values of light-on time, may have reflected ineffective energy use for lighting in the offices. Apart from the group with occasional participation, the lack of significant difference in mean values of occupancy time between the self-reported- and logged data suggests that the diary form could also reflect ineffective use.
Table 7 shows an example of energy used for lighting by manual controls (1) calculated using mean values of light-on time from the logged data and mean values of light-on time as well as occupancy time from the self-reported data. The energy use was compared to the mean values for light-on time from the logged data and the differences are presented in Fig. 2 and 3. The figures show overestimation and underestimation of the energy use when using the mean values from the self-reported data, respectively.
TABLE 7. Example of lighting energy use calculated using mean values from the self-reported-and logged data Group of Occupants Data Used for Mean Energy Use Calculation Value (kWh/Day) (h) All occupants Light-on time 6.68 0.51 from the logged data Light-on time 7.00 0.53 from the self-reported data Occupancy time 5.29 0.40 from the self-reported data Occupants with full Light-on time 6.81 0.52 participation from the logged data Light-on time 7.00 0.53 from the self-reported data Occupancy time 4.69 0.36 from the self-reported data Occupants with occasional Light-on time 6.51 0.49 participation from the logged data Light-on time 7.00 0.53 from the self-reported data Occupancy time 6.12 * 0.46 from the self-reported data * Significant difference in mean values between the self-reported- and logged data. Fig. 2. Percentage of difference in calculated energy use between mean values for light-on time from the self-reported data and for light-on time from the logged data. Light-on time from the self-reported data All 4% Full participation 2% Occasional participation 8% Note: Table made from bar graph. Fig. 3. Percentage of difference in calculated energy use between mean values for occupancy time from the self-reported data and for light-on time from the logged data. * Significant difference in mean values between the self-reported and logged data. Occupancy time from the self-reported data All -22% Full participation -31% Occasional participation -6% * Note: Table made from bar graph.
6 METHODOLOGICAL CONSIDERATIONS
We were able to show that the diary form can be a suitable and reliable means of measuring lighting use in offices where lighting controls as well as orientation varied among the offices and where the occupants had different work tasks. However, the study could not examine the reliability of self-reported data on regulating desk lamps and adjusting window blinds and curtains due to the limitations of the electronic measurements.
Our findings also show that the diary form was somewhat unsuccessful in obtaining dimming activities as well as in reflecting the amount of energy used by dimming controls. Regarding this, the form could include dimming activities indicating the specific lighting levels. Moreover, the design of the diary form could be improved to obtain more accurate responses and to capture factors influencing lighting use. For example, it could provide opportunities for office occupants to freely report their movements and reasons when adjusting the light. To further capture behaviors, experiences as well as other psychological factors in relation to lighting use in work places, event-based Ecological Momentary Assessment (EMA) approaches [Shiffman and others, 2008] can be applied by diaries. Besides paper-and-pencil diaries, electronic diaries collecting data through new technologies such as smart phones could be done less obtrusively, especially with a large sample size.
The study period of one year made it possible to identify two groups of participants according to the extent of their participation in the self-reported part, though the number of participants was small. This could help to evaluate the diary form for short-and long-term uses. With the small sample size, however, personal factors (such as age, gender and work task) influencing the extent of participation could not be identified. To use the forms once every two-months may also be questionable regarding the suitability for monitoring the occupants' behavioral patterns.
7 CONCLUDING REMARKS
From our results, it appears that the diary approach can be a cost- and time-effective method for studying lighting use at work places. It can be especially advantageous when a large sample size is required and when a nonintrusive method is needed to obtain data from participants. The concluding findings are:
(i) The diary forms were able to determine the occupants' switching patterns in their offices during the day.
(ii) The diary forms were more successful in measuring occupants' use of general lighting than their time present in the offices.
(iii) The diary forms were able to reliably measure light-on time and could therefore be employed to determine lighting energy use.
(iv) The diary forms are more reliable for measuring the use of lighting and determining energy use in short- than in long-term studies.
The findings above should, however, be confirmed by examining the diary form with a larger sample size and for different types of offices and organizations since different environmental and social contexts may affect the suitability and reliability of the form.
Additional benefits of the use of diary forms could be achieved by analyzing data both on individual and group levels. The diary forms could also be used in open-plan offices where occupants may share general lighting. Moreover, office occupants may become more aware of the effects of their behaviors on energy use by self-reports. According to Palm and Ellegard , the data of self-reported activities were useful both for reflecting households' energy use and providing feedback. Given this, the diary form could be a pedagogic tool providing feedback both to organizations and individuals in interventions where office occupants are to be motivated to change their behaviors.
The authors thank the Swedish Energy Agency for funding the study presented in this paper.
(1) Energy use = wattage x time; the figure of wattage, 75.87 Wh/h [Maleetipwan-Mattsson, 2012] was applied.
Bell PA, Greene TC, Fisher JD, Baum A. 2001. Environmental psychology, 5th ed. Fort Worth (TX): Psychology Press. 634 p.
Boyce PR. 1980. Observations of the manual switching of lighting. Lighting Res Technol. 12(4):195-205.
Boyce PR, Veitch JA, Newsham GR, Jones CC, Heerwagen J, Myer M, Hunter CM. 2006. Occupant use of switching and dimming controls in offices. Lighting Res Technol. 38(4):358-378.
Chao YL, Lam SP. 2011. Measuring responsible environmental behavior: self-reported and other-reported measures and their differences in testing a behavioral model. Environ Behav. 43(1):53-71.
Dubois MC, Blomsterberg A. 2011. Energy saving potential and strategies for electric lighting in future North European, low energy office buildings: a literature review. Energ Buildings. 43(10):2572-2582.
Escuyer S, Fontoynont M. 2001. Lighting controls: a field study of office workers' reactions. Lighting Res Technol. 33(2):77-96.
Galasiu AD, Newsham GR, Suvagau C, Sander DM. 2007. Energy saving lighting control systems for open-plan offices: A field study. Leukos. 4(1):7-29.
Gatersleben B, Steg L, Vlek C. 2002. Measurement and determinants of environmentally significant consumer behavior. Environ Behav. 34(3):335-362.
Horelli L. In Bechtel RB, Churchman A, editors. 2002. Handbook of environmental psychology. A Methodology of participatory planning. p 607-628. New York (NY): John Wiley and Sons. 722 p.
Hunt DRG. 1979. The use of artificial lighting in relation to daylight levels and occupancy. Build Environ. 14(1):21-33.
Jennings JD, Rubinstein FM, DiBartolomeo D, Blanc SL. 2000. Comparison of control options in private offices in an advanced lighting control testbed. J Illum Eng Soc. 29(2):39-60.
Love JA. 1998. Manual switching in private offices. Lighting Res Technol. 30(1):45-50.
Maleetipwan-Mattsson P. 2012. Evaluation of lighting control systems: A case of occupants' experience and use of different control systems in office environments. Lund (Sweden): Lund University, Department of Architecture and Built Environment. Environmental Psychology Monograph No. 21 2012. 64 p.
Maniccia D, Rutledge B, Rea MS, Morrow W. 1999. Occupant use of manual lighting controls in private offices. J Illum Eng Soc. 28(2):42-56.
Masoso OT, Grobler LJ. 2010. The dark side of occupants' behaviour on building energy use. Energ Buildings. 42(2):173-177.
Moore T, Carter DJ, Slater AI. 2003. Long-term patterns of use of occupant controlled office lighting. Lighting Res Technol. 35(1):43-59.
Olsen ME. 1981. Consumers attitudes toward energy-conservation. J Soc Issues. 37(2):108-131.
Pallant J. 2007. SPSS Survival manual a step by step guide to data analysis using SPSS for Windows. 3rd ed. Berkshire (UK): Open University Press. 335 p.
Palm J, Ellegard K. 2011. Visualizing energy consumption activites as a tool for developing effective policy. International Journal of Consumer Studies. 35(2):171-179.
Reinhart CF, Voss K. 2003. Monitoring manual control of electric lighting and blinds. Lighting Res Technol. 35(3):243-260.
Scherbaum CA, Popovich PM, Finlinson S. 2008. Exploring individual-level factors related to employee energy-conservation behaviors at work. J Appl Soc Psychol. 38(3):818-835.
Shiffman S, Stone AA, Hufford MR. 2008. Ecological momentary assessment. Annu Rev Clin Psychol. 4:1-32.
Siero FW, Bakker AB, Dekker GB, vandenBurg MTC. 1996. Changing organizational energy consumption behaviour through comparative feedback. J Environmental Psychol. 16(3):235-246.
Starik M, Marcus AA. 2000. Introduction to the special research forum on the management of organizations in the natural environment: a field emerging from multiple paths, with many challenges ahead. Academy of Management Journal. 43(4):539-546.
The Swedish Energy Agency. 2007. Forbattrad energistatistik for lokaler- "Stegvis STIL" Rapport for ar 1 inventeringar av kontor och forvaltningsbyggnader. ER. 2007:34. 92 p.
Tso GKF, Yau KKW. 2003. A study of domestic energy usage patterns in Hong Kong. Energy. 28(15):1671-1682.
Veitch JA, Gifford R. 1996. Assessing beliefs about lighting effects on health, performance, mood and social behavior. Environ Behav. 28(4):446-470.
Pimkamol Maleetipwan-Mattsson (1) *, Thorbjorn Laike PhD (1), and Maria Johansson DSc (1)
(1) Department of Architecture and Built Environment, Lund University, Box 118, SE-221 00, Lund, Sweden.
* Corresponding author: Pimkamol Maleetipwan-Mattsson, E-mail: firstname.lastname@example.org
|Printer friendly Cite/link Email Feedback|
|Author:||Maleetipwan-Mattsson, Pimkamol; Laike, Thorbjorn; Johansson, Maria|
|Date:||Apr 1, 2013|
|Previous Article:||User preferences in office lighting: a case study comparing Led and T5 lighting.|