by

Only people that are Female, 0-17 years old, and indicate via the "Fertil" field that they have had children, will be included in the first analysis.

Note that the census bureau does not provide data items that explicitly identify any of the household level categorizations that we have chosen. They have been created by our tools.

We are not going to reference the household level section here, but it is available here for those who might be interested.

Line numbers have been added to the analysis for easy reference. Looking at the person level section of the analysis, we notice that:

1. Line 1 shows us that 916 teen mothers are the only mother in their household, and no children (under 5) live with them.
2. Line 8 shows us that 925 teen mothers live with another mother in their household, and no children (under 5) live with them.
3. Line 11 shows us that 969 teen mothers live with another mother in their household, and one child (under 5) lives with them, but the teen mother is not a parent in a subfamily. We cannot tell who the child belongs to from this. We wonder if the children (and/or the non-teen mothers) are in subfamilies.
4. Line 13, by far the largest cell, shows us that 2112 teen mothers live with another mother in their household, and one child (under 5) lives with them, and the teen mother is a parent in a subfamily.
5. Lines 15 and 17 show us that 320 and 284 teen mothers live with another mother in their household, and two children (under 5) live with them. In one case the teen mother is a parent in a subfamily, and in the other case she is not. Again, we wonder if the children are in subfamilies, which would allow us to determine which mother is the parent.

Our first question is answered by line 17 of this second analysis. We were wondering about the 969 teenage mothers that were not in subfamilies - were the children in the households in subfamilies? Line 17 tells us that 698 of the children are not in subfamilies, so they likely belong to to the teenage mothers. The remaining 271 children likely belong to the other mother in the household.

The next question concerned the 320 and 284 teen mothers that lived in households with two young children, and one other mother. Who do these children belong to? Line 25 shows us that 561 children were not in a subfamily, and likely belong to the 320 teen mothers that were not in a subfamily. Similarly, line 30 shows us that 532 children were in a subfamily, and likely belong to the 284 teen mothers that were in a subfamily.

A few new questions are arising. Some of our numbers aren't coming out as perfectly as we'd like them to. For example, in households that have a single child in subfamily 1, we expect to see at least as many parents in those households in subfamily 1. See lines 20 and 22. We have 2486 parents with 2749 children, when we expect to see a larger number in the "parent" category than in the "child" category. This is probably an artifact of the census bureau's weights, but to be sure, we'll run another analysis. In the new analysis, we'll add a column for "People", which is an unweighted person count.

(Note: In the previous analyses, only households with at least 1 teen mother were included. The following analyses include all households in PA, split out so that they can be differentiated. That is why the following tables are longer.)

The new person level analysis helps by bringing the numbers a lot closer, but it doesn't completely solve our problem. We can see from line 116 that 112 people are behind the parent count of 2486. Line 118 tells us that 115 people are behind the child count of 2749. These are families with only one child - how can there possibly be more children than parents?

To answer this one, we run yet another person level analysis. This time we create a new household categorization to categorize each household based on the number of parents in parent/child subfamilies (labeled "Parents"). Line 208, for example, shows us that there are households which have children in subfamilies that do not have parents in subfamilies. This demonstrates how it is possible for certain segments to have more total children than parents in one-child subfamilies. And this explains our final question. (Click here to see the household level analysis.)