Alcohol and Parenting: The Effects of Maternal Heavy Drinking

3. Methodology

The Canadian National Longitudinal Study of Children & Youth (NLSCY), Cycle l, affords a data set with which to investigate maternal drinking and the effect on her offspring. The specific methodology for the survey and the rationale for the determination of questions are detailed in the NLSCY Project Team's overview statement (NLSCY Project Team, l995). The first cycle contains four questions dealing with alcohol consumption:

During the past l2 months, have you/he/she had a drink of beer, wine, liquor or any other alcoholic beverage?
During the past l2 months, how often did you/he/she drink alcoholic beverages (7 ratings possible from daily to less than once a month)?
How many times in the past l2 months have you/he/she had five or more drinks on one occasion? and
In the past l2 months, what is the highest number of drinks you/he/she had on one occasion?

The stated reason for the inclusion of these questions was "because of the potential impact on adults' physical or mental health, the family's economic situation, and family relationships" (NLSCY Project Team, l995).

The brevity and nature of these four questions present a number of restrictions for any subsequent analyses. First, currently the "field" defines the abuse of alcohol in terms of the consequences of drinking rather than the frequency/quantity of alcohol consumed. Additionally, questions appropriate to previous or current definitions of dependency are lacking. Thus, direct application of the results to the growing abuse/dependency literature is somewhat problematic. Second, in analyzing drinking, traditionally and importantly, one would also want to compare quantity consumed by frequency of drinking while also considering dosage, the time taken to consume the drink, and the weight and sex of the individual. The survey does not define "a drink," a certain quantity per drink, the weight of the subject or the time course when drinking. Also unknown is whether individuals who may not have consumed alcohol in the past l2 months were previously problem drinkers. Inspite of these concerns, the question "how many times in the past l2 months have you had 5 or more drinks on one occasion?" does permit a rough estimate of the frequency per year that blood alcohol level exceeded 0.08%. Conrod et al. (l997) have demonstrated that this estimate correlates highly with other self-report estimates of alcohol consumption, frequency of estimates of intoxication, and laboratory volume of alcohol consumed. Further the measure of frequency per year that blood alcohol level exceeded 0.08% has been shown to be successful in discriminating between the presence and absence of problem drinking symptoms (Conrod et al., l997). Hence, the definition of severity of drinking selected for analyzing the NLSCY data was drinking per se and the frequency of consuming five or more drinks on one occasion.

A priori, four groups were formed, non drinkers, light drinkers, moderate drinkers, and heavy drinkers. The non drinkers were those who had not consumed alcohol in the past l2 months. The light drinkers, for women, were individuals who had never consumed more than 5 drinks per occasion and for men, for whom data were also available, those who had done so on less than 7 occasions. Moderate drinking women were defined as having less than l2 occasions of five or more drinks while for men it was 7 to 52 occasions. Heavy drinkers for women were those who had more than l2 occasions per year of 5 or more drinks and for men more than 52 occasions. These groupings were based on the results from the study by Conrod and colleagues (l997), in which men were characterized as heavy drinkers if they consumed 5 or more drinks on more than 52 occasions per year. Further, as the National Institute of Mental Health Epidemiologic Catchment Area study found that alcohol abuse/dependence is considerably more prevalent among men than among women, by a factor of about 5 to l (Helzer and Pryzbeck, l988), we attempted to approximate this same male/female ratio of heavy drinkers in the present study by characterizing women as heavy drinkers if they consumed 5 or more drinks, on more than l2 occasions per year. For the characterization of light drinkers, the number of times per year men or women would consume 5 or more drinks on one occasion needed to reflect quantities of alcohol which would most likely not intoxicate women, and rarely intoxicate men. Lastly, the moderate drinker group was defined by default, being those who did not meet inclusion criteria for either heavy or light drinking groups.

Although there does appear to be some support for our classification of drinker groups, alcohol researchers have not agreed on operational definitions for "Light", "Moderate", and "Heavy" alcohol consumption. In a review by Abel and Kruger (l995), examples were given to illustrate the great disparity in the operational definition of these terms: "Moderate" consumption has been variously described as l-l3 drinks/week (Virji, l99l), 4-l3 drinks/week (Williams and DeBakey, l992), 5-l0 drinks/week (Sulaiman et al., l988), l-7 drinks/week (Thorogood et al., l993), and 7-20 drinks/week (Rostand et al., l990). "Heavy" drinking has been characterized as l0-l2 drinks/week (Sulaiman et al., l988), 8 or more drinks/week (Thorogood et al., l993), l4 or more drinks/week (Williams and DeBakey, l992), 2l or more drinks/week (Rostand et al., l990), or more than 28 drinks/week (Greeley et al., l993).

In the NLSCY survey 89.l% of the questionnaires were completed by the children's biological mothers. Hence, we decided to analyze only those surveys completed by the biological mother and for the most part focused on her drinking behaviour. Table l presents descriptive data for the four groups for the mothers who completed the survey. The age, number and percentage of women fitting each group as well as similar data for their spouse are presented.

Table 1 : Descriptive Data on Sample
	Mean Number of Children in Household	Mother			Spouse
	Mean Number of Children in Household	Number	%	Mean Age	Number	%	Mean Age
Non Drinkers	2.16	2475	21	32.6	2134	21.5	35.6
Light Drinkers	2.03	5766	48.6	33.4	5078	51	36.2
Moderate Drinkers	1.92	2650	25.6	31.2	2443	24.6	34.4
Heavy Drinkers	1.93	366	3.5	31.5	274	2.7	35.5

These results support the classification of groups employed in this study. The proportion of women heavy drinkers for example draws support from other studies. The National Institute of Mental Health Epidemiologic Catchment Area Program estimated the one-month prevalence US rate of alcohol abuse/dependence to be 2.8% in noninstitutionalized adults between l980-l984, and 4.7% for 6 month prevalence (Regier et al., l988). Although data for the latter study was collected about l5 years ago, surveys in the US and Canada conducted over the past 2 decades have found little evidence of major changes in drinking levels or drinking problems among women in general (Wilsnack and Wilsnack, l99l). Further, a report from the National Comorbidity Survey (Kessler, et al. l994) showed the l2 month prevalence for men for alcohol abuse was 3.4% and for dependency was l0.7% while for women it was respectively l.6% and 3.7%.

Other studies support some aspects of the proportions in each of the drinker groups of the present investigation. For example, in a study by Lipton (l994), 928 men and women of the Los Angeles Epidemiological Catchment Area were grouped into 23.8% abstainers, l3.l% light, 22.0% light moderate, 23.8% moderate, and l7.2% heavy drinkers. In a prospective study (Wannamethee and Shaper, l997) of cardiovascular disease involving 7735 men ages 40-59, 9.7% were non- or ex-drinkers, 29.7% were occasional drinkers, 37.l% were light drinkers, l9.3% were moderate drinkers, and 4.l% were heavy drinkers. Furthermore, the proportion of heavy drinkers in the present study is also similar to that of Caracci (l992), who found that 3.2% of female subjects reported they were heavy drinkers. Lastly, the proportion of abstainers in the present study also replicates that found by Slicker (l997), where l9.6% of a sample of university students were abstainers. Notably, when compared to percentages of individuals with dependency problems in society, it would appear that our heavy drinking category is perhaps even more stringent than typical psychiatric diagnostic criteria.

The disproportionate number of individuals in the various drinking categories generally obviated against using a continuous analytic procedure throughout the variables. Where possible, analysis of continuous data was completed and results generally conformed to the ascertained group differences reported below. Appendix 1 lists the variables against which drinking level is compared. Variables were analyzed under the headings: "Family Characteristics", "Maternal Health", "Maternal Ratings of Child's Behaviour", "Teacher Ratings", and "Child's Self-Report". In addition, 4 and 5 year olds completed the Peabody Picture Vocabulary Test; school-aged children completed mathematical abilities tests; 10 and 11 year olds completed a self-report questionnaire. The data of one child per mother was randomly selected for analysis so as to avoid the presence of a family bias predicated on the number of children. All analyses were done on weighted means which corrected for possible sampling biases. That is, the obtained sample was corrected to conform with actual representations as reflected in the census. In addition, because of the large number of variables (Appendix 1) and subsequent analyses, a stringent criterion of significance of p<.001 should be considered. All other findings should be viewed as trends.

Contents

Alcohol and Parenting: The Effects of Maternal Heavy Drinking - October 1998

3. Methodology