5. Family Environment

Vorige Start Omhoog Volgende

Table 11 
Results of Statistical Control on CSA-Symptoms Relations

 

Reports of self-perceived negative effects, especially when of a lasting nature, certainly suggest that CSA can cause harm for some persons with certain types of CSA experiences. The issue we focus on here, however, is whether CSA typically causes harm. Previously we saw that CSA was statistically significantly correlated with poorer adjustment, although the magnitude of this relation was small.

One of the most fundamental principles in scientific methodology is that correlation is not causation. That is, for example, just because race is correlated with IQ, that does not mean that race causes differences in IQ. It could be that some third variable, such as home environment or socioeconomic status, is responsible for the race-IQ association.

To illustrate this concept, let’s take this simplistic example, shown in Figure 1. As you go from small towns to small cities, to large cities, the number of churches will increase. Further, as you go from small towns to big cities, the amount of crime also increases. Does this mean that building new churches will increase crime, or tearing some down will decrease crime? No, because there’s a third variable, population, that is responsible for both. As population grows, more churches are built and more crimes occur. If we factored out population size in this example, the correlation between number of churches and amount of crime would probably disappear.

We examined the relationship between CSA and symptoms using this idea. In this case, the third variable that might be causing both is family environment. A broken home, or one containing physical abuse or emotional neglect, could predispose children or teenagers to counternormative behavior, such as using drugs or engaging in sexual activities that are classified as CSA. A broken home could also impair their adjustment. In this way, the relationship between CSA and symptoms that we found in our meta-analyses could be the result of family environment, rather than the CSA experiences.

From our previous meta-analyses, we know that for college subjects CSA accounted for 0.81% of the adjustment variability. We conducted a series of meta-analyses to determine what percent of the variability in CSA was accounted for by family environment. The result was 1.69%. We next conducted a series of meta-analyses to determine what percent of the adjustment variability was accounted for by family environment. The result was 8.41%. In other words, these results show that family environment was substantially more important in terms of being able to account for adjustment variability than CSA was -- by a factor of 10.

These results also show that CSA was indeed confounded with family environment--those who had CSA tended to come from poorer, more disorganized family settings. These findings together suggest that the statistically significant, but small relationship between CSA and adjustment may not be causal.

Thirteen of the college studies used statistical techniques to factor out, or hold statistically constant, family environment, when examining the relationship between CSA and adjustment (see Table 11). The 14 samples from these studies examined 83 CSA-adjustment relations. Before statistical control, 41% of these relations were statistically significant. After statistical control--that is, after removing the effects of family environment--only 17% were statistically significant. This represents a 59% reduction. Since CSA-adjustment relations within a given study tend to be correlated, we computed the percent reduction in statistical significance by using one overall result per study. Computed this way, the reduction in statistically significant results rose to 83%. These findings strongly support the possibility that many instances of statistically significant associations between CSA and adjustment are spurious. In particular, these findings argue against the popular assumption that CSA typically causes harm.

Table 11

Results of Statistical Control on CSA-Symptoms Relations

Study

Type of control

Significant results

N

Before

After

% reduction

Brubaker, 1999

Separated categories

1

1

0

100

Cole, 1988

Hierarch. Regression

5

3

0

100

Collings, 1995

ANCOVA

10

8

6

25

Fromuth & Burk, 1989, mw

Hierarch. Regression

13

6

6

0

Fromuth & Burk, 1989, se

Hierarch. Regression

13

0

0

-

Fromuth, 1986

Hierarch. Regression

13

4

1

75

Gidycz et al., 1995

Path analysis

3

0

0

-

Greenwald, 1994

Hierarch. Regression

1

0

0

-

Harter et al., 1988

Path analysis

2

1

0

100

Higgins & McCabe, 1994

Hierarch. Regression

2

2

0

100

Lam, 1995

Multiple regression

3

0

0

-

Long, 1993

Multiple regression

2

1

0

100

Pallotta, 1992

ANCOVA

13

6

0

100

Yama et al., 1992

ANCOVA

2

2

1

50

Totals

83

34

14

59a

Note. N indicates the number of symptom measures whose relation to CSA status was examined (or was intended to be by the study authors) by using statistical control. "Before" indicates the number of relations significant before applying statistical control; "After" indicates the number of significant relations after applying statistical control. "Reduction" indicates the percent of significant relations that became nonsignificant after statistical control.

a Based on the percent of total significant relations that became nonsignificant after control. The unweighted percent reduction was 83%.

Vorige Start Omhoog Volgende