melodygd
请问: 也就是,如果抽样的客群总体在某属性上无差异的话,其中的部分客群是否还会在该属性上存在显著差异?
比如,对一类客户群体做性别的差异分析,发现该客户群中没有性别的差异。那么,是否有必要对该类客群根据其他属性做进一步细分,看细分后的各客群是否存在性别差异?
请高人指点!多谢!
friend
我认为应该没有必要再做了,即使是有差异也只是个别很少的属性里有。我也是新手,仅作参考,还请指教!
anita_jiu
[quote]引用第0楼melodygd于2007-06-29 15:21发表的“关于客户细分 差异检验的疑问”:
请问: 也就是,如果抽样的客群总体在某属性上无差异的话,其中的部分客群是否还会在该属性上存在显著差异?
比如,对一类客户群体做性别的差异分析,发现该客户群中没有性别的差异。那么,是否有必要对该类客群根据其他属性做进一步细分,看细分后的各客群是否存在性别差异?
.......[/quote]
某属性 - what element exactly, pls? Is it related to demographical elements (e.g. income, age) or something else, please (e.g. purchase behaviour)? Would you mind giving more details about your sample or research, please? Otherwise it is difficult to discuss your question on a sound ground.
There are arguments for using homogeneous sample; then of course, there are also debates against this application. It depends on your research nature. If you may explain a bit more, perhaps it would be helpful for others to interact your query more appropriately.
melodygd
多谢anita_jiu的提醒。 我的具体问题是: 对一类客户群体做满意度在性别方面的差异分析。做方差分析后,发现该客户群的满意度在性别上没有差异。如果我对该类客群根据其他属性(比如年龄)做进一步细分,那么,细分出来的各个客群的满意度在性别上有可能存在差异吗?
简单的说就是, 对于在某个个人属性上同质的客户群体,在按照其他一些个人属性细分为更小的客户群体后,这些客户群体在原来那个个人属性上可能存在差异吗?
不知说没说清楚。还请高人指点。
anita_jiu
[quote]引用第3楼melodygd于2007-06-30 20:33发表的“”:
多谢anita_jiu的提醒。 我的具体问题是: 对一类客户群体做满意度在性别方面的差异分析。做方差分析后,发现该客户群的满意度在性别上没有差异。如果我对该类客群根据其他属性(比如年龄)做进一步细分,那么,细分出来的各个客群的满意度在性别上有可能存在差异吗?
简单的说就是, 对于在某个个人属性上同质的客户群体,在按照其他一些个人属性细分为更小的客户群体后,这些客户群体在原来那个个人属性上可能存在差异吗?
不知说没说清楚。还请高人指点。[/quote]
Thanks for the details. I think that nothing (in general) would stop you playing with your data. This means, yes, certainly you may continue to try to identify any difference exists in your sample through elements such as age, gender, education, income, family status etc. However, I'd like to point out one thing: so what? This is a key question to answer. Have found or haven't found difference, so what? If statistical difference exists, does it have any impact on the practical side? Does it help marketers and researchers better understand your particular market?? Any implications for marketers and/or researchers?
By the way, more questions to come: in general, is your sample a homogeneous sample (i.e. share commonalities) or not, please? How big is your sample? The nature of your sample also affects statistic analysis outcomes. I'd suggest you review your research objectives at this point.
gzw6681467
按照某个标准(如性别)分组的的组间没有差异,但按另一标准(如年龄、教育程度、个性、地区等)进行分组,组间是可以存在差异的。所以为了把数据挖掘得更深些,把数据信息挖掘得更全面,我们通常要从多个角度对数据进行分析。