Overlap of Opinion Leadership across Consumer Product Categories
Overlap of Opinion Leadership across Consumer Product Categories
复制标题
不同消费产品类别的意见领袖重叠
DOI：
10.1177/002224377100800221
复制
发表时间：
1971
影响因子：
6.1
通讯作者：
S. Sudman
中科院分区：
文献类型：
--
作者：
S. Sudman
矛盾就在前两种说法之间。如果在不同产品类别中的意见领导力很高，这将限制意见领导者的数量。然而，研究结果似乎有力地证实了第一个结论。第二种说法没有数据支持，第三种说法可以解释原因。如果将6个产品类别中5个或6个类别的6.6%的意见领袖排除在外，那么产品类别之间的领导重叠将大大减少。下表根据金-萨默斯文章表3的修正数据，给出了意见领导力成对重叠的实际和预期百分比，省略了这一群体。虽然包装食品与家用清洁剂和洗涤剂之间，女性服装时尚与化妆品和个人美容辅助品之间，以及大型和小型电器之间仍然存在很强的成对相关性，但大多数其他差异，即使显着，也不是很令人印象深刻。另一种快速查看的方法是King-Summers文章中表4中的相关矩阵。在那里，忽略5或6个类别中6.6%的受访者将使解释方差R2减少0.066，r减少0.26。因此，从表4中的数字中减去0.26给出了剩余成对相关性的估计值。从这些结果中，很难推断出任何强有力的证据来证明产品之间的高意见领导重叠。任何关于此类一般意见领袖的证据都必须取决于6.6%的受访者，他们报告自己是5或6个类别的意见领袖。
The contradiction is between the first two statements. If opinion leadership over across product categories is high, this would limit the number of opinion leaders. However, the results presented seem to strongly confirm the first conclusion. It is the second that is unsupported by data, and the third statement may explain why. If the 6.6% of the respondents who reported they were opinion leaders in 5 or 6 of the 6 product categories are set aside, then the leadership overlap across product categories is substantially reduced. The table gives the actual and expected percentages for pairwise overlap of opinion leadership based on the revised data from Table 3 of the King-Summers article, omitting this group. While there are still strong pairwise correlations between packaged food products and household cleansers and detergents, between women's clothing fashions and cosmetics and personal grooming aids, and between large and small appliances, most of the other differences, even if significant, are not very impressive. Another quick way to see this is in the correlation matrix in Table 4 of the King-Summers article. There, omitting the 6.6% of the respondents in 5 or 6 categories would reduce the explained variance R2 by .066 and r by .26. Thus, subtracting .26 from the figures in Table 4 gives an estimate of the remaining pairwise correlations. From these results, it is difficult to deduce any strong evidence for high opinion leadership overlap across products. Any evidence of such general opinion leaders must then depend on the 6.6% of the respondents who reported they were opinion leaders in 5 or 6 categories.