Centering the dependent variable

Is it acceptable to group-mean center the dependent variable? My dependent variable is work-family conflict and I have four different ethnic groups. I'm concerned that members of one ethnic group consistently report more conflict due to cultural bias in answering the questionnaire. If I group mean center the dependent variable in each ethnic group, then this bias will be eliminated. I'm concerned that this may create other statistical problems. Help!