Printer Friendly

Social Media Applications Preference by Generation and Gender: An Exploratory Study.


Millennials or generation Y are a growing segment of the US population that grew up in an online socially-networked world. Generation X grew up during the early stages of the Internet, and have been enjoying a period of a socially-connected world. Baby Boomers grew up before the Internet age during the mainframe age, and the early introduction of the Mac and the IBM PC Jr. According to (Strauss & Howe, 1991), Millennials are those individuals born between 1982 and 2004. Millennials have surpassed Baby Boomers as the largest living generation in the US in 2015 (Fry, 2016). In 2016, Millennials are those of age between 18 & 29, Generation X are between 36 & 51, and Baby Boomers are between 52 & 70 (Fry, 2016). According to a 2012 survey by the Pew Research Center, and cited in a report by Consumer Behavior in 2013, the following percentages of consumers use social networks: Millennials (18-29): 86%; Generation X: (30-49): 72%; Younger Boomers (50-64): 50% and; Seniors (65+): 34%.

In the context of social media usage by different groups of generations, the majority of studies found during the last 10 years focused on Millennials. This is consistent with the goals of social media applications providers and marketers, since the Y generation represents the segment of current and future users and consumers. For millennials, there is a gender difference in the intentions to support charitable events online through Facebook. Females have higher intentions to support events than males (Paulin et al., 2014). In a study of the influence of positive and negative product-related messages through social media such as Facebook, Twitter, and MySpace, results showed that Millennials were significantly influenced by product comments regardless of the sender, family member or more distant person (Sago, 2010). However, according to a Forbes magazine article in 2015, based on interviews of 1,300 Millennials, only 1% said that a compelling advertising would make them trust a brand more (Schawbel, 2015). Men and women have different expectations from social networking, such as Facebook and Twitter. Also, they use them differently (Clipson et al., 2012). In a study of the intention to use mobile applications during a convention, there was no significant difference in information seeking motivation among the three generation groups; also, LinkedIn and Facebook were the most popular for all groups (Lee & Lee, 2014). However, there was significant difference between generation X and Baby Boomers for future intentions to use mobile applications.

A widely used tool in the area of data mining is the association rule mining, also called market-basket analysis (Agrawal et al., 1993). It is a proven technique for discovering meaningful patterns in database transactions. It has been used to identify trends in data and predict future events. Association is the discovery of correlations among a set of items in a transaction. An association rule of the form: A [right arrow] B implies that if A occurs then B is likely to occur. Also, IF {set of values [V.sub.i]'s} THEN {set of value [V.sub.j]'s}. The model for mining association rules employs the support-confidence framework (Agrawal et al., 1993).



The data used in this study were downloaded from the datasets available through the Pew Research Center ( This data set, collected November 2010, includes more than 2000 records and uses at least 25 variables. During preprocessing, six variables were selected. Four variables relate to the usage of Twitter, Facebook, MySpace, and LinkedIn. The other variables are for gender and age. Then, all records with missing values were eliminated. Finally, the variable age was utilized to create new variables to indicate the generation using the scheme adopted by the Pew Research Center. Note that since this dataset was collected in November 2010, therefore, the age bracket for each generation was reduced by six years, using the scheme suggested by the Pew Research Center. Table 1 provides a snapshot of the resulting dataset which totals 739 records.


Let I = {[I.sub.i]} i =1, 2 the set of two input variables representing the generation and the gender. Let O = {[O.sub.j]} j =1, 2, 3, 4, the set of four output variables representing the social media applications: Twitter, Facebook, LinkedIn, and MySpace. Each survey case is also called a record or an item set or a transaction. This item set can be represented by two groups of values {LHS, RHS}, where the left hand side (LHS) represents the generation and the gender as input I, and the right hand side (RHS) represents the social media applications as output O. The format {LHS} [right arrow] {RHS} constitutes a rule: IF {LHS} THEN {RHS}. Using the support-confidence algorithm (Agrawal et al., 1993), we can compute the support of the LHS as: supp(LHS) = frequency of a combination of generation and gender in the database and confidence of LHS and RHS as conf(LHS [right arrow] RHS) = supp(LHS U RHS) / supp(LHS). The most widely used measures for rule interestingly are support and confidence. But, for the purpose of this study, we only display the confidences values for the fifteen rules obtained. The strategy used to identify the rules was to limit the LHS to the possible combinations of generation and gender. The RHS could include any combination of the four social media applications. This resulted in fifteen rules shown in Table 2 below.

For Table 2, note the following: Gen for Generation: Millennials (Y), Generation X (X), and Baby Boomers (BB); Sex for Gender: Male (M), and Female (F). Also, due to the low values for confidence, and to provide information about the frequencies, the confidence values are shown as ratios.

As shown on Table 2 below, the confidence values for rule # 13 are 0 for all generations and genders. Therefore, there will be 14 rules representing 14 categories of social media preferences. For the next step in the analysis, the original data have been updated according to the rules or categories shown on Table 2. Each of the 739 transactions or records was assigned to 1 of the 14 preference categories. Then, a series of statistical ANOVA tests were performed. Finally, the possibility of building a neural network (NN) was investigated. For this NN, the input neurons generation and gender were used to predict the output preference category. The results of the data analysis are discussed in the next section.


As shown on figure 1 below, there is a significant interaction effect between social media preference category and generation ([F.sub.2, 736] = 8.247, p < 0.001). This result should be expected due to the chronological sequence for the introduction of the different social media applications. For example, since generation Y users were exposed to Facebook at an early age, they are more likely to use it than the Baby Boomer generation users. Also, as shown in figure 2, there is also a significant interaction effect between social media preference category and gender ([F.sub.314, 423] = 1.37, p < 0.01). However, the difference between Millennials and Baby Boomers was more significant ([F.sub.226, 276] = 1.517; p < 0.001) than that between generations Y and X ([F.sub.226, 234] = 1.251; p = 0.044). However, there was no significant difference between the social media applications preferences between generation X and Baby Boomers.

Anova: Single Factor

Groups     Count  Sum   Average    Variance

ClusterY   227    2141   9.431718  8.272972
ClusterX   235    2337   9.944681  6.608038
ClusterBB  277    2873  10.37184   5.451813

Source of Variation   SS       df      MS         F      P-value

Between Groups        110.269    2  55.13449   8.246633  0.000287
Within Groups        4920.673  736   6.685697
Total                5030.942  738

Source of Variation  F crit

Between Groups       3.007959
Within Groups


F-Test Two-Sample for Variances

                     Variable 1  Variable 2

Mean                   9.866667   10.00708
Variance               8.071338    5.893567
Observations         315         424
Df                   314         423
F                      1.369517
P(F<=f) one-tail       0.001326
F Critical one-tail    1.187939

Figure 3 below shows the results of building a neural network to predict the social media preference category using generation and gender as inputs. Out of the 739 cases, 503 were used as a training set, and the remaining cases were used to test the network. The preliminary results show a promising lead with an accuracy of 62% in training, and 59% in testing.

                       Model Summary
Training  Cross Entropy Error            705.839
          Percent Incorrect Predictions  37.6%
          Stopping Rule Used             1 consecutive step(s) with no
                                         decrease in error (a)
          Training Time                  0:00:00.50
Testing   Cross Entropy Error            357.565
          Percent Incorrect Predictions  41.1%

                     Case Processing Summary
          N         Percent

Sample    Training  503        68.1%
          Testing   236        31.9%
Valid     739       100.0%
Excluded    0
                    Total     739

Dependent Variable: Cat_of_Pref
(a). Error computations are based on the testing sample.


The purpose of this study was to investigate differences in the social media applications preferences among Millennials, generation X, and Baby Boomers, as well as between male and female users. The results demonstrated clear differences between Millennials and Baby Boomers, as well as between male and female users. Also, the preliminary results of a neural network were promising. The limitations to this research include the exclusion of other demographic attributes such as income, ethnicity, and educational level. In addition, future research should consider comparing users within the same generation by segmenting a generation further according to age and other criteria.


Agrawal, R., Imielinski, T., & Swami, A. (1993). Mining association rules between sets of items in large databases. ACM SIGMOD International Conference on Management of Data, Washington DC.

Clipson, T., Wilson, A., & DuFrene, D. (2012). The social networking arena: Battle of the sexes. Business Communication Quarterly, 75(1), pp. 64-67.

Fry, R. (2016). Millennials match baby boomers as largest generation in U.S. electorate, but will they vote? The Pew Research Center report, May 16.

Lee, S. & Lee, C. (2014). An exploratory study of convention specific social media usage by attendees: Motivations and effect of generation on choice of convention information source and intention to use mobile applications. Journal of Convention & Event Tourism, 15, 135-149.

Lim, J-S, Lim, K-S., & Heinrichs, J. (2014, fall). Gender and mobile access method differences of millennials in social media evaluation and usage: An empirical test. Marketing Management Journal.

Paulin, M. Ferguson, R., Schattke, K., & Jost, N. (2014). Millennials, social media, prosocial emotions, and charitable causes: The Paradox of gender differences. Journal of Nonprofit & Public Sector Marketing, 26, 335-353.

Sago, B. (2010, fall). The influence of social media sources on millennial generation consumers. International Journal of Integrated Marketing Communications, 7-18.

Schawbel, D. (2015). 10 New findings about the millennial consumer. Forbes, Jan 20.

Strauss, W., & Howe, N. (1991). The history of America's future, 1584-2069. New York: Harper Collins.

Moncef Belhadjali, Norfolk State University

Sami M. Abbasi, Norfolk State University

Gary L. Whaley, Norfolk State University

       Millennials     Generation X  Baby Boomers  Total
       (Generation Y)  30-45         46-64

Male    92             109           114           315; 43%
Female 135             126           163           424; 57%
Total  227; 31%        235; 32%      277; 37%      739


Rule#  {LHS} [right arrow] {RHS}

 1     {Gen&Sex} [right arrow] {Twitter}
 2     {Gen&Sex} [right arrow] {Facebook}
 3     {Gen&Sex} [right arrow] {LinkedIn}
 4     {Gen&Sex} [right arrow] {MySpace}
 5     {Gen&Sex} [right arrow] {Twitter&Facebook}
 6     {Gen&Sex} [right arrow] {Twitter&LinkedIn}
 7     {Gen&Sex} [right arrow] {Twitter&MySpace}
 8     {Gen&Sex} [right arrow] {Facebook&LinkedIn}
 9     {Gen&Sex} [right arrow] {Facebook&MySpace}
10     {Gen&Sex} [right arrow] {LinkedIn&MySpace}
11     {Gen&Sex} [right arrow] {Twitter&Facebook&LinkedIn}
12     {Gen&Sex} [right arrow] {Twitter&Facebook&MySpace}
13     {Gen&Sex} [right arrow] {Twitter&LinkedIn&MySpace}
14     {Gen&Sex} [right arrow] {Facebook&LinkedIn&MySpace}
15     {Gen&Sex} [right arrow] {Twitter&Facebook&LinkedIn&MySpace}

Rule#  Y,M    Y,F     X,M     X,F     BB,M    BB,F

 1      3/92   2/135   1/109   2/126   1/114    2/163
 2     49/92  82/135  56/109  82/126  63/114  121/163
 3      0/92   0/135   9/109   2/126  11/114    6/163
 4      4/92   3/135   3/109   1/126   4/114    3/163
 5      9/92   4/135   9/109   9/126   2/114    4/163
 6      0/92   0/135   0/109   0/126   1/114    0/163
 7      0/92   0/135   0/109   0/126   2/114    0/163
 8      7/92   9/135  15/109  13/126  15/114   10/163
 9     11/92  23/135   7/109  13/126   6/114    9/163
10      0/92   0/135   1/109   0/126   1/114    0/163
11      3/92   3/135   4/109   2/126   3/114    4/163
12      4/92   5/135   3/109   0/126   2/114    2/163
13      0/92   0/135   0/109   0/126   0/114    0/163
14      2/92   0/135   1/109   2/126   3/114    1/163
15      4/92   4/135   0/109   0/126   0/114    1/163
COPYRIGHT 2016 American Society for Competitiveness
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2016 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Belhadjali, Moncef; Abbasi, Sami M.; Whaley, Gary L.
Publication:Competition Forum
Date:Jan 1, 2016
Previous Article:Competitiveness and Sustainability in a Changing Global Landscape: The Business Strategy Game Simulation.
Next Article:The Adoption of Mobile Game Applications as Sustainable Social and Behavioral Change.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters