Methods of Communication Research and Statistics

Open the data file: MediaGedrag.sav, which investigates young people's media-related behaviour. We are now going to investigate whether, in addition to time spent on the internet, number of Facebook friends is also influenced by frequency of reading books and watching television.

a. Run a regression analysis for the 'time spend on the internet' (independent variable) and 'number of Facebook friends' (dependent variable). What is the value of the standardised regression coefficient (beta) (report according to APA6)?

b. Run a regression analysis using 'frequency of reading books' as the independent variable and number of Facebook friends as the dependent variable. What is the value of the standardised regression coefficient?

Answer: b* = -.77

c. Run a regression analysis using 'frequency of watching television' as the independent variable and number of Facebook friends as the dependent variable. What is the value of the standardised regression coefficient?

d. When you compare the results of a, b, and c, which independent variable has the strongest influence on the 'number of Facebook friends' variable?

Frequency of watching television

e. Calculate again the effect of 'time spend on the internet' on 'number of Facebook friends', but now keep 'frequency of reading books' and 'frequency of watching television' constant. In other words, run a multiple regression analysis where you enter all three independent variables into the 'independent' field. Is the partial effect of 'time spend on the internet' on 'number of Facebook friends' the same as the simple effect of 'time spend on the internet' as established under a? If not, is it stronger or weaker? What is the value?

0.26 is lower than 0.82. The partial effect (b* = 0.26) is weaker than the simple effect (b* = 0.82).

f. Is the partial effect of 'frequency of reading books' on 'number of Facebook friends', when 'time spend on the internet' and 'frequency of watching television' are kept constant, the same as the simple effect of 'frequency of reading books' established under b)? If not, is it stronger or weaker? What is the value?

-0.02 is weaker than -0.77. The partial effect (b* = -.02) is much weaker than the simple effect (b* = -.77).

g. Is the partial effect of 'frequency of watching television' on 'number of Facebook friends', when 'time spend on the internet' and 'frequency of reading books' are kept constant, the same as the simple effect of 'frequency of watching television' established under c)? If not, is it stronger or weaker? What is the value?

0.73 is weaker than 0.94. The partial effect (b* =.73) is indeed weaker than the simple effect (b* =.94).

h. Which of the three independent variables is the strongest predictor of 'number of Facebook friends'? How can you tell?

Frequency of watching television; it has the highest standardized regression coefficient

i. Does the conclusion you gave in h) correspond to the conclusion you gave in d)? If not, why not?

j. What is the intercept of the multiple regression analysis (with the three independent variables)? What does this mean?

27.24, someone who does not use the internet does not watch television and does not read books has 27.24 Facebook friends

k. What is the value of the unstandardised regression coefficient of the 'book' variable and what does this mean?

-0.03, when time spend reading a book increases by one unit, number of Facebook friends decreases by 0.03, while keeping the other independent variables 'time spend on the internet' and 'watching television' constant

l. Is the effect of reading books on number of Facebook friends significant? Report you results.

Minutes spend reading books is not a significant predictor of number of Facebook friends, b* = -0.02, t = -0.08, p = .939, 95% CI [-0.80, .0.75]

m. What is the value of the unstandardised regression coefficient of the 'internet' variable and what does this mean?

0.18, when time spend on the internet increases by one unit, the number of Facebook friends increases by 0.18, while keeping the other independent variables 'reading books' and 'watching television' constant

n. Is the effect of time spend on the internet on number of Facebook friends significant? Report your results.

Minutes spend reading books is not a significant predictor of number of Facebook friends, b* = 0.26, t = 1.28, p = .249, 95% CI [-0.16, 0.52]

o. What is the value of the unstandardised regression coefficient of the 'television' variable and what does this mean?

0.40, when time spend watching television increases by one minute, number of Facebook friends increases by 0.40, while keeping the other independent variables 'reading books' and ‘time spend on the internet' constant.

p. Is the effect of time spend watching television on number of Facebook friends significant? Report your results.

Minutes spend reading a book is a significant predictor of number of Facebook friends, b* = 0.73, t = 2.84, p = 0.030, 95% CI [-0.56, 0.74]. The more television young people watch, the more Facebook friends they have, when reading books and time spend on the internet are kept constant.

90.2%, the variance in time spend on the internet, reading books, and watching television combined explains 90.2% of the variance in the number of Facebook friends.

r. Is the whole model significant? State the null hypothesis for the whole model and provide the relevant results.

time spend watching television, time spend on the internet and reading books do not explain the variance in number of Facebook friends in the population, F (3, 6) = 18.35, p = .002, R2 = .90

SPSS exercise 6.4