We tested possible differences from the website, geographic region, and you may ethnicity using t-assessment and you may data off difference (ANOVA) for my review here the LIWC class proportions. On the two websites, half a dozen of several t-examination was in fact significant about following the classes: first-people one [t(3998) = ?5.61, p Additional Table 2 to possess form, simple deviations, and you will contrasts ranging from ethnic communities).
Contrasts shown high differences between White as well as almost every other ethnic organizations within the four of your own half dozen tall ANOVAs. Thus, we provided ethnicity as the good dummy-coded covariate inside the analyses (0 = White, step 1 = Any kind of cultural groups).Of several ANOVA testing associated with geographic area, simply a few was in fact high (family and self-confident feelings). Once the variations just weren’t commercially significant, i don’t think geographical region from inside the next analyses.
Volume of keyword explore is obvious in the detailed statistics (find Dining table 1) and via keyword-clouds. The term-affect strategy illustrates the essential widely used words over the entire shot as well as in each one of the a long time. The definition of-affect system immediately excludes particular conditions, plus posts (a good, and, the) and you can prepositions (so you can, having, on). The rest posts terms and conditions was scaled in size in line with their regularity, undertaking an intuitive portrait of the most prevalent posts words around the the new take to ( Wordle, 2014).
Profile step one shows the fresh 20 common stuff terms and conditions included in the whole test. As well as rise above the crowd, probably the most frequently used words was like (looking inside the 67% regarding pages), particularly (searching in the 62% from pages), appearing (looking inside the 55% away from profiles), and you may some body (looking during the 50% away from users). Therefore, the most popular terms and conditions had been equivalent round the a long time.
Figure dos reveals next 29 most typical posts terminology inside the brand new youngest and oldest age groups. By removing the initial 20 common blogs conditions along side attempt, i illustrate heterogeneity on the relationships users. Next 29 terminology on the youngest generation, high level percentage terms and conditions integrated score (36% off users on youngest age group), go (33% from profiles on youngest age group), and you can performs (28% of pages from the youngest generation). Having said that, the newest earliest generation had large rates of terms such as take a trip (31% off users regarding earliest age group), great (24% of users regarding the oldest age group), and you will matchmaking (19% regarding profiles from the oldest age bracket).
Second 30 most common words on the youngest and you will earliest many years groups (just after subtracting brand new 20 most commonly known conditions away from Shape step one).
To check on hypotheses, the brand new percentage of terminology about dating character that suit for each and every LIWC category supported because depending variables within the regressions. I checked out age and you can gender given that separate parameters and changing to own web site and you may ethnicity.
Hypothesis step 1: Old decades would be from the a higher part of words regarding pursuing the categories: first-person plural pronouns, family, relatives, wellness, and you will positive feelings.
Findings largely supported Hypothesis step one (look for Table 2). Five of your four regressions shown a serious head effect to have many years, in a way that due to the fact ages of new reputation copywriter enhanced, the fresh new part of terminology throughout the classification improved throughout the pursuing the categories: first-individual plural, nearest and dearest, health, and you may confident emotion. We discovered zero extreme years effect to your proportion off conditions on family members category.
a Gender: 0 (female) and step one (male). b Webpages: The two other sites had been dictomously coded due to the fact step 1 and you can 0. c Ethnicity: 0 (White) and you can step 1 (Ethnic or racial minority).
an excellent Gender: 0 (female) and you can step one (male). b Webpages: The 2 other sites was dictomously coded once the step 1 and 0. c Ethnicity: 0 (White) and step one (Cultural otherwise racial fraction).
