{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

C what general underlying construct do you think each

Info iconThis preview shows page 3. Sign up to view the full content.

View Full Document Right Arrow Icon
c. What general underlying construct do you think each of these three items is measuring? I think that these three items are measuring the confidence of people in certain government factions. d. Let’s try constructing the Cronbach’s alpha. This measure is a scale reliability coefficient. The higher the value the more consistently the three items scale together. Generally values of greater than 0.6 or so are considered minimally acceptable while values greater than 0.7 are considered good. Use the following command alpha conpol congov conciv. . Paste the output into the output sheet . Do these measures have an acceptably high scale reliability coefficient? These measures do indeed have a high scale reliability coefficient of 0.7096. PART 3: Constructing the scale Now let’s try constructing the scale. The simplest way to do this is to use the gen command to generate a new variable and set it equal to the sum of the three variables from which our scale is to be composed. We can do this quite simply: gen conscale = conpol+congov+conciv. This command tells stata to generate a new variable called conscale that is equal to the sum of the values from conpol, congov and conciv . Tab this new variable and paste the output into the output sheet . You might want to try listing the first ten observations of the scale and the variables from which it is composed to get a sense of how this worked. Now let’s try another version of an additive scale. Rather than simply adding the scales together, let’s construct a measure that consists of a count of the number of institutions in which the respondent has confidence. Run the following code in your do-file. gen conscale2=0 replace conscale2=conscale2+1 if conpol>2 replace conscale2=conscale2+1 if congov>2 replace conscale2=conscale2+1 if conciv>2 What is this code doing? How is it different than what we ran to generate the first conscale ? The code is replacing variables and showing the real changes made, so it is very different compared to the fist conscale and generates more real change.
Background image of page 3
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

Ask a homework question - tutors are online