Relationships

Internal associations of socioeconomic scale components

Table 1 shows the correlation coefficients of associations within the various socioeconomic scale components, region wide. The two poverty scores, which are developed from some of the same source data, are the most closely related components.

TABLE 1: Coefficients of correlation between components of socioeconomic scale (Pearson Correlation coefficients)

	Education	FCPI	Tenure	Poverty	Poverty Intensity	Employment

Education	1.0000	-.0883	-.0044	-.2780	-.1766	.4040
# cases	130	130	130	130	130	130
Signif	.	.318	.961	.001	.044	.000

FCPI	-.0883	1.0000	-.1385	.5581	.3972	-.3934
# cases	130	130	130	130	130	130
Signif	.318	.	.116	.000	.000	.000

Tenure	-.0044	-.1385	1.0000	-.1673	-.1255	.1220
# cases	130	130	130	130	130	130
Signif	.961	.116	.	.057	.155	.167

Poverty	-.2780	. 5581	-.1673	1.0000	. 9057	-.6042
# cases	130	130	130	130	130	130
Signif	.001	.000	.057	.	.000	.000

Pov Intensity	-.1766	.3972	-.1255	.9057	1.0000	-.5915
# cases	130	130	130	130	130	130
Signif	.044	.000	.155	.000	.	.000

Employment	.4040	-.3934	.1220	-.6042	-.5915	1.0000
# cases	130	130	130	130	130	130
Signif	.000	.000	.167	.000	.000	.

FCPI = Families with children receiving public assistance income.
# cases = Number of cases evaluated, Significance = Level of two-tailed significance.

The numbers in this and the following two tables focus on the nature and strength of relationships between different measures of well-being; for example, poverty and tenure or socioeconomic status and community capacity. A relationship is said to exist between two variables when a change in the value of one variable is accompanied by a change in the value of the other variable. For example, in the Klamath region communities with high poverty rates tend to have higher unemployment rates than those with low poverty rates. Understanding the nature of these relationships is an important step towards social explanation and prediction. It should be noted, however, that while correlation may aid in the prediction of social condition it does not imply causation.

Correlation analysis measures both the direction and the strength of the relationships between two variables. The resulting correlation coefficients can be used to compare the strength of relationships between multiple pairs of variables. In this study two different correlation measures are used to measure two basic types of data. Pearson's product moment correlation coefficient is used to measure relationships of interval data. This includes all the socioeconomic data as well as the continuous socioeconomic scale. Spearman's rank-order correlation coefficient is used with ordinal data. The primary ordinal data in this study is the capacity score, which is a scale of five discrete values from "very low" to "very high." Also, the socioeconomic score is reduced to a seven point ordinal scale for presentation purposes.

While Pearson's correlation coefficient and Spearman's rank-order coefficient are used with different data types, they are interpreted similarly. The direction of a relationship is indicated by the sign of the coefficient. A positive coefficient indicates a positive relationship. That is, when variable X increases, variable Y also tends to increase. A negative sign indicates an inverse relationship where variable Y tends to decrease as variable X increases. The strength of a relationship is indicated by the value of the correlation coefficient. A coefficient of 1.00 indicates a perfect or pure relationship. That is, the change in variable X is exactly the same as the change in variable Y. Likewise, a zero value indicates that the two variables are completely independent. In this case, knowledge of one variable indicates nothing about the value of the other. So, the closer that a coefficient is to a value of 1.00 the stronger the relationship. Generally, coefficients less than 0.10 are considered to indicate the lack of any significant relationship.

Since the interpretation of correlation analysis generally involves an assessment of the coefficient value, the number of records in the analysis (n), and the results of a significance test, these three values are presented together. Significance is indicated as "p" in reports of Pearson coefficients and as "sig" with Spearman rank order coefficients. Significance testing indicates the likelihood that the measured correlation of a sample set is representative of a larger population or set of values. The larger the sample relative to the total population, the less likely are the results to have occurred purely by chance. In this study, however, correlation analysis is always used on the total population (the 180 block group aggregations in the Klamath region or a subset in subregional analyses) rather than a sample set. Significance testing then, is largely irrelevant in this study.

Table 2 shows the Pearson correlation coefficients resulting from an analysis of the relationships between the socioeconomic scale and the individual components of the scale for the entire Klamath region and for aggregations within each of the six subregions. As would be expected, since the scale is based on these components, there is a relatively strong association between each component variable and the socioeconomic scale. Although the poverty score has one of the highest correlations with the socioeconomic scale on average, the scale is not determined by any single component.

TABLE 2: Coefficients of correlation between socioeconomic scale and scale components for the Klamath region by subregion (Pearson Correlation coefficients)

	Region	Subregion
	Klamath	Rogue	North Coast	Trinity	Modoc Plateau	Central Siskiyou Corridor	Northern Sacramento Valley
Tenure	.4406	.3842	.2915	.6184	-.0182	.7190	.7828
# cases	130	40	34	11	17	13	15
Signif	.000	.014	.094	.043	.945	.006	.001

Poverty Intensity	-.7048	-.7729	-.7371	-.7913	-.5420	-.2357	-.8698
# cases	130	40	34	11	17	13	15
Signif	.000	.000	.000	.004	.025	.438	.000

Poverty	-.8043	-.8284	-.8244	-.8663	-.7295	-.5783	-.8799
# cases	130	40	34	11	17	13	15
Signif	.000	.000	.000	.001	.001	.038	.000

Education	.5388	.5066	.7293	.0203	.5072	.3682	.7188
# cases	130	40	34	11	17	13	15
Signif	.000	.001	.000	.953	.038	.216	.003

Employment	.7907	.7334	.8361	.7538	.6885	.7153	.7374
# cases	130	40	34	11	17	13	15
Signif	.000	.000	.000	.007	.002	.006	.002

FCPI	-.6590	-.6057	-.6915	-.3763	-.6267	-.7783	-.8346
# cases	130	40	34	11	17	13	15
Signif	.000	.000	.000	.254	.007	.002	.000

FCPI = Families with children receiving public assistance income.
# case = Number of cases (aggregations) evaluated, Signif = Level of two-tailed significance.

Socioeconomic status and capacity

Socioeconomic status and capacity are both important components of well-being, but measure different aspects of it. Correlation analysis between the two measures for the study region reveals a positive but weak relationship between socioeconomic status and community capacity (Spearman rank order coefficient 0.2621, n=130, sig=0.003). This relationship remains positive at subregional levels, but the strength of the relationship varies considerably with correlation coefficients ranging from a low of 0.0541 (n=40, sig=0.740) in the Rogue subregion to a high of 0.5201 (n=11, sig=0.101) in the Trinity subregion. Table 3 lists Spearman rank order coefficients between the socioeconomic score and the community capacity score within each subregion. While human capital is partially reflected in the socioeconomic scale through educational attainment and income-related components, social and physical capital are not.

TABLE 3: Spearman rank order coefficients between socioeconomic score and community capacity score by subregion

Subregion	Spearman Correlation Coefficient	Number of Cases	2-tailed Significance
Rogue	.0541	40	.740
North Coast	.4818	34	.004
Trinity	.5201	11	.101
Modoc Plateau	.4482	17	.071
Central Siskiyou Corridor	.0921	13	.765
North Sacramento Valley	.0878	15	.756

Table 4 shows the juxtaposition of capacity and socioeconomic status scores for the 130 aggregations. Aggregations with medium-low to very low capacity (1–2) and a very low to medium-low socioeconomic status scale score (1–3) are those considered to have the lowest level of well-being. A total of 25 aggregations, or 19.2 percent of all aggregations, fall into this group, which constitutes only 8.1 percent of the total regional population.

Aggregations with high and very high socioeconomic status (6–7) and with a capacity score of medium or higher (3–5) are viewed as having the highest level of well-being. Eleven aggregations, 8.5 percent of all aggregations, constituting four percent of the regional population, falls into this group. Low capacity associated with high socioeconomic status is not, in general, associated with a low well-being. This is because the residents of aggregations with high socioeconomic status can and often do "buy" their way out of difficulties that others must work internally to overcome. Nonetheless, even among the aggregations with high socioeconomic status, capacity itself is a component of well-being.

The remaining aggregations, with moderate to moderately high well-being, can be further divided into three groups. Aggregations with medium to high capacity (3–5) and very low to medium-low socioeconomic status (1–3) have a moderate level of well-being. A total of 8.5 percent of all aggregations fall within this group. Similarly, the 23.8 percent of aggregations with low to medium-low capacity (1–2) and medium to medium-high socioeconomic capacity (4–5) also have a moderate level of well-being. While the former group has a lower socioeconomic score, the higher capacity suggests a greater ability to take advantage of opportunities than the latter group of aggregations, which has a higher socioeconomic score. The group of aggregations with medium to high capacity (3–5) and medium to medium-high socioeconomic status (4–5) and the group with very low to medium-low capacity (1–2) and high to very-high socioeconomic status (6–7) has a moderately high level of well-being. This group makes up 40 percent of all aggregations and represents 66 percent of the population within the Klamath region.

It is important to point out that the combination of a high capacity rating and a high socioeconomic status score does not mean that all residents of an aggregation enjoy a high level of well-being (though they are more likely to than if the aggregation had a low capacity and very low socioeconomic status score). Just as some families may enjoy a considerably higher level of well-being than others in the same aggregation, some groups-ethnic, occupational, or other-may collectively have considerably lower well-being.

TABLE 4: Number of aggregations by capacity and socioeconomic status score

		CAPSCORE
SESCORE		1	2	3	4	5	Total	percent
1		1	2				3	2.3
2		5	6	2	1		14	10.8
3		4	7	4	2	2	19	14.6
4		10	10	17	11	6	54	41.5
5		4	7	5	6	2	24	18.5
6		2	2	5	4	1	14	10.8
7			1			1	2	1.5
Total		26	35	33	24	12	130
percent		20.0	26.9	25.4	18.5	9.2	100.0
Shading indicates varying levels of well-being based on capacity and socioeconomic status
	Low levels of well-being
	Moderate levels of well-being
	Moderately high levels of well-being
	High levels of well-being

Single-parent households

Twenty-six percent of families with children under 18 years of age in the Klamath region are single parent families. All but three aggregations (Trinity Lake, Castella/La Moine/South Dunsmuir, Honeydew/Petrolia) have some single-parent families, with the percentage ranging from six percent to 51 percent. Over 77 percent of all single parent families in the region are headed by a female. At the aggregation level, however, this rate ranges from as low as 27 percent to 100 percent, except in three aggregations (Hornbrook/Hilt, Johnson Park, Scotia) where all single parent families are headed by a male. Male single parent families only exceed female single parent families in ten out of 127 aggregations with single parent families. The North Coast subregion has the highest percentage of single-parent households (29 percent).

FIGURE 8: Poverty rates of family households with children by family type and subregion

For the entire Klamath region, single parent households with children are far more likely to have incomes below the poverty line than are married couple households with children. Single male-headed households with children are over twice as likely as married couple households to have incomes below the poverty level, while single female-headed households with children are over five times as likely. Figure 8 shows subregion level poverty rates of female-headed households with children, male-headed households, and family households headed by married couples.

Correlation analysis indicates an inverse relationship between socioeconomic status and single-parent families (Pearson coefficient -.5136, n=130, p=0.000), female-headed single parent families (-.5077, n=130, p=0.000), and to a lesser degree, those single parent families headed by a male (-.1962, n=130, p=0.025).

Spatial characterization and relationships

Based on the isolation scale, the aggregations of the Modoc Plateau and Trinity subregions are, on average, the most isolated. These areas also have some of the most isolated aggregations in the region. The isolation scale scores of five aggregations within these subregions are more than two standard deviations above the mean. These include Hoaglin/Kettenpom/Zenia/Lake Mountain, Hyampom, and Salyer in Trinity and Davis Creek/New Pine Creek and Cassel/Hat Creek/Old Station in the Modoc Plateau. Seventy percent of the aggregations in the Modoc Plateau have moderately high, high or very high isolation scale scores. Only Klamath Falls has a low or moderately low isolation scale score. Seventy three percent of the aggregations in Trinity have moderately high, high or very high isolation scale scores. The aggregations of the Rogue and Northern Sacramento Valley subregions are, on average, the least isolated. Fifty-three percent of the aggregations in each of these subregions have moderately low or low isolation scale scores. On average, the aggregations of the North Coast also have relatively low isolation scores. Nonetheless, the North Coast subregion has the greatest variation in isolation scale scores of any subregion and also has some of the most isolated aggregations in the Klamath region including Hoopa Valley Indian Reservation, Honeydew/Petrolia, Orleans, and Ettersburg/Shelter Cove/Whitethorn. Twenty-six percent of the aggregations in the North Coast subregion have moderately high, high or very high isolation scale scores. Figure 9 shows the percentage of aggregations within each subregion by relative isolation score.

Socioeconomic status tends to decrease as aggregation isolation increases, but this relationship is variable. Correlation analysis indicates a weak inverse relationship between isolation and the socioeconomic scale in the Klamath region (Pearson coefficient -.3189, n=130, p=0.000). The relationship between isolation and the socioeconomic scale varies at the subregional level, but is positive in all but one subregion. The inverse relationship is strongest in the Trinity (Pearson coefficient -.5813, n=11, Sig=0.061), Siskiyou Corridor (Pearson coefficient -.4874, n=13, Sig=0.091), and the North Coast (Pearson coefficient -.3936, n=34, Sig=0.021) subregions. It is weakest in the urbanized Northern Sacramento Valley subregion (Pearson coefficient -.0626, n=15, Sig=0.825). There is actually a positive relationship between the isolation and socioeconomic scales in the Modoc Plateau subregion (Pearson coefficient 0.3583, n=17, Sig=0.158). The 24 aggregations with high to very high isolation scores have an average socioeconomic score of 3.4, compared to an average of 4.5 for the 28 aggregations with low isolation scale scores.

FIGURE 9: Percentage of aggregations by relative isolation score by subregion

Similarly, community capacity tends to decrease as isolation increases, but this relationship is also variable. There is an inverse relationship between isolation and capacity in the Klamath region (Spearman coefficient -.3109, n=130, Sig=0.000). The relationship between isolation and capacity varies widely at the subregional level. The inverse relationship is strongest in the Northern Sacramento Valley (Spearman coefficient -.8558, n=15, Sig=0.000), North Coast (Spearman coefficient -.5824, n=34, Sig=0.000), and Trinity (Spearman coefficient -.4201, n=11, Sig=0.198) subregions. The relationship is negligible in the remaining three subregions. The 24 aggregations with high to very high isolation scores have an average capacity score of 2.4, compared to an average of 3.5 for the 28 aggregations with low isolation scale scores. All 12 of the aggregations in the Klamath region with high capacity scores have low or moderately low isolation scale scores.

Population, population density, and relationships

The relative population density of aggregations varies considerably across the Klamath region from 0.29 persons per square kilometer in Davis Creek/New Pine Creek to 209.02 persons per square kilometer in Grant Pass. The ten aggregations with population densities over two standard deviations from the mean include some of the most populated aggregations such as Redding, Medford and Arcata, as well as some mid-sized aggregations such as Redwood Area, Burney and Fields Landing/King Salmon. With a population of 927, Scotia is the smallest of these high population density aggregations.

Community capacity tends to be higher as aggregation population density increases. The average capacity score of those aggregations with higher than average population density is 3.4, nearly one point higher than the average capacity score for those with less than average population densities. Also, correlation analysis indicates a positive relationship between aggregation population density and capacity across the Klamath region (Spearman coefficient .3534, n=130, Sig=0.000). This relationship remains positive at the county level although the strength of it varies. It is strongest in the Northern Sacramento Valley subregion (Spearman coefficient .8046, n=15, Sig=0.000), but also high in the North Coast (Spearman coefficient .5523, n=34, Sig=0.001), Central Siskiyou Corridor (Spearman coefficient .4882, n=13, Sig=0.091), and Trinity (Spearman coefficient .4154, n=11, Sig=0.204) subregions.

Given the close relationship between capacity and population among the aggregations of the Klamath region there is also a positive correlation between capacity and total aggregation population (Spearman coefficient .3931, n=130, Sig=0.000). This relationship is weakest in the Rogue subregion (Spearman coefficient .2661, n=40, Sig=0.097), and strongest in the Sacramento Valley (Spearman coefficient .7223, n=15, Sig=0.002).

There appears to be little relationship between population (or population density) and socio-economic status.