Hello again Yahya,

I would recommend using an ordinal logistic regression to make full use of your data. For the independent variables, it may be of benefit to explore a Principal Components/Factor analysis to come up with a set of variables that most represent the variability in your data.

In which software?

Spss

Regardless of what software you use, assuming you can write some simple if-then type programming commands, the idea is as follows: for each variable, you create binary variables (0 or 1), for each of the categories -- except for one category of answers you identify as the 'baseline' group.

For example, for your income variable, since you have four categoires, you create three dummy variables (let's say for the top three income levels, with the bottom level as the baseline). Each of those dummies (call them d_i2, d_i3, d_i4), will take on the value of 1 or 0 based on the response you have for your predictor.

that's a start, at least...

Many thanks Nora for your valuable reply

A dummy variable is a variable for which all cases falling into a specific category assume the value of 1 and all cases not falling into that category assume a value of zero.

A dummy variable is a variable for which all cases falling into a specific category assume the value of 1 and all cases not falling into that category assume a value of zero.

1- In SPSS, Transform => Re-code è Recode into Different Variables è Type the name of the Output variable (ex. Dumsouth) and click change.

2- Click “Old and New Values” è Enter the old value of South (3) and the new value (1) è Add

3- Click “All other values” and then make the New Value (zero) è add

4- Click “System-or user-missing” on Old Value and “System-missing” on New Value è add.

5- Continue è Ok doky…

In other words, dummy variables are used to examine group differences.

Thus, for the variable region:

South would be 1 and all other regions would be zero. When we regress the attitudes toward abortion on the dummy variable "south" we are trying to measure the difference in attitudes toward abortion between the people of the south versus everybody else.

The idea is very simple:

Transform a categorical/nominal variable into a binary variable.

Coding convention: 0 for the value that does not have the characteristic and 1 for the value that has the characteristic.

So 1 for women and 0 for men for the variable "dumgender" or "female"

Hi Yahya,

Here is an example using the income variable that you have in your data. You will end up having 4 dummy variables in the data coded as follows:

Income Inc1 Inc2 Inc3 Inc4

Category 1 1 0 0 0

Category 2 0 1 0 0

Category 3 0 0 1 0

Category 4 0 0 0 1

When you use this variable in a regression analysis, the dummy variable for reference category is omitted (soem softwares will do that automatically for you).

I agree with Tabassum in his example. Thank you.

Many thanks Sankalp,Tabassum and Mahfuz, for this a detailed description of how to created dummy variables.

Hi once again, I have another problem regrding binary logistic regression and wanted to share with you, this is as follows:

I used 5-Likert scale to measure my dependent variables (1 strongly agree, 2 agree, 3 neutral, 4 disagree, 5 strongly disagree) and as you aware that using binary logistic regression requires me to create a binary variables from these 5 levels. For example, strongly agree and agree become (category 1), and disagree and strongly disagree become (category 2) and exclude the neutral (although this significantly cuts down my sample size)?.

One more issue in this regard, my questionnaire has 12 variables and each variable is measured by a group of items. For example, I have education (measured by 7 items), Gender inequality (measured by 6 items), trust (measured by 6 items), career development (measured by 7 items) and so on. Now, and in order to create the binary categories, do I need to take the high and the low score of each variables and find the mean and standard deviation, and exclude the cases in the middle? but this again will significantly cuts down my sample size?

Finally, is it better to use multinominal logistic regression instead of binary logistic regression in my situation?

Many thanks

Use weka's logistic regression. It will do the job for you before generating results. You can also use weka's nominal to binary preprocessing tool.

Hello again Yahya,

I would recommend using an ordinal logistic regression to make full use of your data. For the independent variables, it may be of benefit to explore a Principal Components/Factor analysis to come up with a set of variables that most represent the variability in your data.